Closed shentino closed 2 years ago
My guess at this point is something funky involving integer handling.
1.6.19 confirmed bad 1.6.18 confirmed good
I'm going to do a bisect
Commit cdcdf2bb45b75be04dd087333f808b8682768dac reliably triggers the issue.
The LPC code involved is kotaka::/home/Text/lwo/ansi/painter.c which takes a grid of characters and colors and renders an ANSI color coded buffer using nested loops.
Let me know if you need any more information.
A stack trace doesn't seem to be applicable here, what exactly do you need?
Should I attach screenshots and a copy of the LPC code in question?
Before cdcd2fb: \<redacted>
After cdcdf2b: \<redacted>
The only difference is in which version of DGD I compile.
Both screenshots were generated from a cold boot of kotaka.
To be clear, I assumed it was my own code at fault at first, but the fact that changing which commit of DGD I compile seems to be the only way to trigger this combined with it having worked before seems to rule out my own code as a culprit.
I made sure not to change anything in the LPC between the two compilations of DGD.
...that's weird, your "how to reproduce" comment doesn't seem to appear in the issue itself.
I'll whip up a reproduction kit and instructions soon.
On Mon, Feb 14, 2022 at 12:00 AM Felix A. Croes @.***> wrote:
How to reproduce?
— Reply to this email directly, view it on GitHub https://github.com/dworkin/dgd/issues/69#issuecomment-1038767326, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMHOIASULU6QEZIMZ7NODU3CZDRANCNFSM5OKHDLUQ . You are receiving this because you authored the thread.Message ID: @.***>
Your request for reproduction instructions came through just fine via email. For some reason your comment didn't show up in the issue itself.
Stand by, this is going to take a bit.
There's a few different pieces of game data involved that I'll need to track down and include.
Should be fixed by https://github.com/dworkin/dgd/commit/9d4805c0198cf12385750c14eae0f1042cc64dc3.
Fix confirmed
For some reason I can reliably get corrupted ansi output from connecting to my test mud when I use the newly compiled dgd, but not a previously compiled version.
I tested this locally and confirmed.
No fatal signal so I can't give a stack trace.
I'm going to see if I can finger the triggering commit.